The r(AI)cial Divide of Big Tech and Data Center Buildout

Anthony Karefa Rogers-Wright

14 Oct 2026

🖨️ Print Article

Protesters march against xAI in Memphis. Stu Boyd II - USA TODAY Network/Reuters

Artificial intelligence and hyperscale data center buildout constitute a new front of environmental racism and a project to optimize white supremacy, with Black communities targeted as sacrifice zones.

The specter of a cross-class white “supremacy” consolidation is more pronounced than ever in the United States (U.S.). All signs point to a death spiral of U.S. empire, with the deterioration of already pernicious material conditions further exacerbated by the accelerated atrophy of basic social systems, the expansion of an iniquitous economic system that is pushing more and more working class people to the margins of penury, while invisibilizing and sacrificing the poor of all races and ethnicities. The racial capitalocene (aka climate change) is increasingly deleterious for critical earth systems and is wiping out entire species, including human beings, while driving more and more to extinction. As these conditions worsen, individualism and competition - both key components of capitalism - for diminishing natural resources will only increase the chances that the specter of a cross-class white “supremacy” consolidation becomes a lived reality.

The indefatigable proliferation of unregulated and unrestrained artificial intelligence (AI) technology coupled with accelerated build-out and operation of hyper-scale data centers points to an irrefutable truth. We are collectively in the epoch of late-stage capitalism at a scale that even V.I. Lenin may not have been able to comprehend. And yet, while all poor and working class people are socially, politically, and economically incarcerated by AI and hyper-scale data center proliferation, there is absolutely a r(AI)cial divide that is prominent in this equation - and Black folk form the denominator.

I have referred to the advent of artificial intelligence and hyper-scale data centers as the Frankenstein of environmental racism and a project of increasing and optimizing the efficiency of white “supremacy” ideology. As it pertains to the Frankenstein reference, Big Tech cartels have effectively vertically integrated many sectors of the U.S. polity to serve its needs. From lawmakers at all levels of government - local, State, and federal - who are members of both corporate and corrupt political parties (Democrats and Republicans), to social services, corporations, and, as I have written previously, even mainstream environmental organizations who have allowed themselves to be purchased and elite-captured by corporate entities. Additionally, the fusion of Big Tech with the biggest polluters in the world, including Big Oil, Big Gas, and Big Nuclear - sectors whose pollution disproportionately impacts Black, Brown, Indigenous, and poor people - forms an environmental, climate, and public health killing monster.

The conglomeration (or the Frankensteinization) of these sectors provide Big Tech cartels with an open invitation to exercise environmental racism and accentuate the notion that Black, Brown, Indigenous, and poor communities are nothing more than energy and economic sacrifice zones that are undeserving of people(s)-centered human rights. This much is made clear as majority-Black communities in cities like Memphis and New Orleans are being subjected to toxic air emissions from the burning of fracked gas to power hyper-scale data centers - and, in some cases, like in Memphis, without federally mandated air permits.

Regarding the increased efficiency of white “supremacy” ideology, in addition to hyper-scale data centers, AI is also powered by Large Language Models (LLMs). Stanford University’s Center for Human-Centered Artificial Intelligence (HAI) defines LLMs as, “an AI system trained on massive amounts of text data to understand and generate human-like language.” HAI adds, “It uses deep learning techniques, specifically neural networks with billions of parameters, to predict and produce coherent text, answer questions, translate languages, write code, and perform various other language-based tasks.” This characterization on the surface seems innocuous and neutral enough; however, a deeper analysis of LLMs reveals that white “supremacy” ideology and racism are firmly ensconced in these systems.

For instance, the Proceedings of the National Academy of Sciences (PNAS), a peer-reviewed journal of the National Academy of Sciences, released a report in September 2024 entitled, Covert Racism is Baked into AI Models. Therein, author Amy McDermot notes:

“Large language models (LLMs), such as ChatGPT, promise easy tools to help users craft essays, make hiring decisions by scanning resumes and cover letters, or, someday, even run background checks by, for example, screening social media posts. The danger in each of these cases is that the models can be biased or even racist due to the training data the LLM employs.”

McDermot also points out, “Racism in AI is nothing new, but most studies of racism in LLMs have uncovered overt examples of prejudice, associating certain stereotypical, derogatory adjectives with explicit mentions of race.” AI, at the end of the day, is about increasing efficiency in carrying out tasks to uphold and optimize existing systems in the pursuit of generating more capital. As McDermot’s findings demonstrate, AI includes increasing the efficiency of and optimization of systems like white “supremacy.” The findings of her study have real-life impacts for Black folk that are only now being confirmed as the U.S. continues to provide Big Tech cartels with all they need and ask for to continue developing a technology that is already putting Black folk, and the entire world, for that matter, at increased risk of full-spectrum dehumanization as well as social and environmental entropy.

In 2023, the McKinsey Institute for Economic Mobility released a report warning AI buildout will likely increase the racial wealth gap in the U.S. with increased speed and efficiency. The report, entitled The Impact of Generative AI on Black Communities, forecasts, “Generative AI has the potential to widen the racial economic gap in the United States by $43 billion each year.” Additionally, the report warns that without correcting long-standing patterns, Black Americans, who, “capture only about 38 cents of every dollar of new household wealth,” will experience an increase in the wealth gap between Black and white households by $43 billion annually by 2045 due to AI development and usage. These findings are further compounded by the adverse impacts AI is already having on preventing Black folk from securing employment.

Last month, Ebony Magazine, in its article, Intelligence and Its Impact On The Black Community, reported, “...racial bias within an AI application screening platform discriminated against African Americans and Asian Americans across 4 million applications, with African Americans having a higher percentage of discrimination by 26%.” The article also warns, “This rejection could spread across other applicant tracking platforms if multiple companies use the same one and store the same data.” At a time when the Black unemployment rate just increased at the fastest rate since the height of the COVID pandemic, the increased use of AI in every sector of society acts as a cudgel that interdicts any semblance of social mobility for Black folk. With the increased use of AI for employment screening, it’s not just that qualified Black folk have less of a chance of getting the job - a condition that’s already in place if their names are too “ethnic”- as much as they need not even apply.

These trends, statistics and projections suggest that AI could also be named AIR, that is, Accelerated Institutional Racism, as its buildout vindicates a portion of the definition of racism offered in the landmark report released by the United Church of Christ (UCC), Toxic Wastes and Race in the United States, “racism is more than a personal attitude; it is the institutionalized form of that attitude.” There is perhaps no better way to describe LLMs that power AI and the resulting impacts to Black and other oppressed and marginalized folk, as it can also be concluded, with warrant, that the classist elements of LLMs are also having disproportionate impacts on poor folk.

It’s both instructive and necessary to review UCC’s Toxic Wastes and Race in the United States, released in 1987, as many of the same patterns persist as they pertain to data center buildout in the 2020s. The major findings of UCC’s report include:

Race proved to be the most significant among variables tested in association with the location of commercial hazardous waste facilities. This represented a consistent national pattern;
Communities with the greatest number of commercial hazardous waste facilities had the highest composition of racial and ethnic residents. In communities with two or more facilities or one of the nation's five largest landfills, the average minority percentage of the population* was more than three times that of communities without facilities (38 percent vs. 12 percent);
In communities with one commercial hazardous waste facility, the average minority percentage of the population was twice the average minority percentage of the population in communities without such facilities (24 percent vs. 12 percent); and
Although socio-economic status appeared to play an important role in the location of commercial hazardous waste facilities, race still proved to be more significant. This remained true after the study controlled for urbanization and regional differences. Incomes and home values were substantially lower when communities with commercial facilities were compared to communities in the surrounding counties without facilities.

Additionally, UCC’s report found:

Three out of every five Black and Hispanic Americans lived in communities with uncontrolled toxic waste sites;
More than 15 million Blacks lived in communities with one or more uncontrolled toxic waste sites; and
Blacks were heavily over-represented in the populations of metropolitan areas with the largest number of uncontrolled toxic waste sites. These areas include: Memphis, Tennessee (173 sites); Cleveland, Ohio (106 sites); St. Louis, Missouri (160 sites); Chicago, Illinois (103 sites); Houston, Texas (152 sites); and Atlanta, Georgia (94 sites).

Given the documented impacts of hyper-scale data centers to air quality, water quality, and public health, it’s not only fair to characterize them as hazardous/toxic waste facilities, it would be irresponsible and unprincipled not to. In fact, Dr. Michael Cork at Harvard University’s T.H. Chan School of Public Health in a recent interview revealed public health impacts associated with hyper-scale data center buildout, stating, “Together with the Piedmont Environmental Council, we completed an analysis of the Vantage Data Center in Loudoun County, Virginia, which uses on-site gas turbines for power in one of the densest data center corridors in the world. We estimated between $53 million and $99 million in annual health damages from associated air pollution—among the largest health-damage estimates produced for a single facility to date.” Dr. Cork added, “In another high-profile case, we looked at the [then] proposed [now operational] xAI Gas Plant in DeSoto County, Mississippi, near Memphis, which includes 41 gas turbines intended to power xAI’s Colossus 2 data center. We projected an increase in regional air pollution and tens of millions of dollars in estimated annual health damages—in an area that already carries a substantial air pollution burden. We estimated that the most impacted people would be those with high social vulnerability, high levels of asthma, and lower median household income.”

The fact that the most impacted people are those with high social vulnerability, high levels of asthma, and lower median household income points to the fact that the majority of these people are Black folk, which is confirmed in the aforementioned Ebony Magazine piece that reported, “The new demand for AI use has led many companies to place AI data centers in areas that seem low-cost yet are predominantly Black.” Ebony continues:

“Over the last six months alone, we’ve seen numerous predominantly Black cities such as Memphis, Tennessee, Covington, Georgia, and Colleton County, South Carolina suffer the consequences of having AI data centers placed in their town. From contaminated drinking water to severe air pollution to power outages, residents' health and livelihoods are at stake in more ways than one.”

That Black communities are being targeted by Big Tech cartels for the placement of their toxic and dehumanizing data centers only shows that after nearly 40 years since the release of Toxic Wastes and Race in the United States, in the words of 1990’s Hip Hop group Nice and Smooth, “Ain’t a Damn Thing Changed.”

When making the connections between the profligate extraction and consumption of natural resources, associated toxic emissions and other environmental externalities, and increased high energy burden due to the documented increase of power bills, along with increased domestic and international militarism, surveillance and command and control of oppressed and colonized Black/African communities, both here and abroad, the r(AI)cial divide of artificial intelligence and hyper-scale data center buildout becomes even more incontrovertible.

The material conditions of Black and other marginalized and oppressed peoples will only worsen with the advent of AI and hyper-scale. The question then becomes, “Who’s going to come and save us?” Perhaps the best way to answer this question is by identifying who isn’t going to come and save us, as the answer is the Black Misleadership Class that has demonstrated its willingness, including that of certain kneegroes, such as elements of the Congressional Black Caucus (CBC), bend the knee and have their heads tapped by corporate entities.

Last week, in a piece for The Lever, journalist Morgan Harper discussed the outsized influence of corporate actors at the the Congressional Black Caucus Foundation’s (CBCF) Annual Legislative Conference. Morgan, reflecting on U.S. Senate candidate Angie Nixon’s umbrage at the profound presence of corporate interests, writes:

“Nixon’s comments highlight the contradiction at the root of the conference, which has long been the largest annual gathering of Black political power. An event that claims to be driving, ‘the political, economic, and social advancement of Black communities worldwide’ depends on the generosity of companies increasingly seen as holding back that advancement, through problematic business practices and political spending that stalls the legislative action that Black communities need.”

This showcasing of the CBCF’s contradictions and capitulations to the very corporations that are worsening the material conditions of Black folk while also enjoining their people(s)-centered human rights further proves that this band of trash panda (raccoon) politicians actively prioritize Black capitalism over Black liberation.

Morgan also noted the presence of Big Tech at the CBCF’s annual event, exclaiming, “I mean, seriously, before reaching the room where Nixon was disparaging Walmart, I walked past a Palantir welcome sign and spotted the Meta Pavilion.” The presence of these corporations is nothing new and no coincidence. According to some reports the CBCF along with the Congressional Black Caucus Institute (CBCI) have a very incestous relationship with Big Tech cartels. Since 2025, the CBCF and CBCI have collected $3.24 million from Amazon, Meta, Google, and Microsoft. Meta, Google, and Amazon, along with utility companies, including Entergy of New Orleans, that are assisting with hyper-scale data center buildout sit on CBCF’s Corporate Advisory Council. This proves that rather than adhering to its mission of, “using the full Constitutional power, statutory authority, and financial resources of the federal government to ensure that African Americans and other marginalized communities in the United States have the opportunity to achieve the American Dream,” the Black Misleadership Class, in this case the CBC, is playing a leading role in the r(AI)cial divide by promoting increased use of AI and hyper-scale data center buildout in service to Big Tech supremacy.

The Black masses must be extremely vigilant now that we know the Black Misleadership Class is actively working with Big Tech cartels to facilitate hyper-scale data center invasions of Black communities despite polling that reveals 66% of Black folk oppose data center buildout where they reside and numerous reports confirming that Big Tech cartels are targeting Black communities for the buildout of their toxic hyper-scale data centers. One tool of Big Tech that the Black Misleadership Class will likely use to seduce poor and working class Black communities subjected to scarcity culture, abject poverty, as well as over-polluted and generally execrable material conditions is so-called Community Benefit Agreements (CBAs). Big Tech cartels have announced their plans to spend billions of dollars on CBAs to mollify and curtail opposition to the build-out of hyperscale data centers that has cost an estimated $260 billion in 2026 alone, with roughly $1 trillion in data center investment facing sustained or meaningful local opposition.

CBAs run the risk of splitting communities apart, as we know that Big Tech will do all they can to handpick Black capitalists in given communities who will be in the room, while those in opposition will be subjected to ostracization. In his latest book, When We Are Kin: The History and Future of Afro-Indigenous Solidarity, Professor Kyle T. Mays explains how Black capitalism is not inclusive as it pertains to wealth distribution, “Wealth, even in Black communities, is not equally distributed; that’s not how capitalism works. A few will be wealthy, and the rest will be working class and poor.” Mays continues, “A Black capitalist is still a capitalist. Black capitalist desires are still capitalist desires. Black wealth will still be hoarded in the hands of a few and not the majority.” In the context of hyper-scale data center buildout, Black capitalism is a form of white “supremacy,” and so-called CBAs are nothing more than a vehicle to deliver capitalism to more places, especially poorer Black communities suffering from scarcity, increased pollution, and dehumanization.

All to say, CBAs are not solutions; they are part of a scheme that will only intensify the r(AI)cial divide of AI and data center buildout while also literally and socially pitting people and communities against each other. This is why it’s disappointing to see groups like the NAACP even entertaining the idea of CBAs as part of their contradictory-named “Stop Dirty Data Centers” campaign when they are fully aware that no amount of money will mitigate impacts to public health nor re-enshrine the self-determination of Black communities once hyper-scale data centers are built in them.

On the other hand, Frederick County, Maryland, recently rejected a $110 million CBA offered by a hyper-scale data center developer. Members of the Frederick County community were first enticed by the CBA that would have included $30 million to renovate a local Elementary School [situated directly next to the proposed data center]; $40 million for a community center and recreation space; $14.5 million for a workforce development training center; $10.5 million for agricultural land preservation; $10 million for berms, plantings and trails; $5 million for a community solar project; and $1 million for a new fire truck for the Carroll Manor company. However, the money came with a major catch - in exchange for the “benefits,” Frederick County would be barred from applying new zoning rules or other regulations for the data center for a period of at least eight years. Upon review, residents of Frederick County understood what was really happening, with one proclaiming in a public comment, "No amount of money will compensate for the destruction data centers will cause to our lives,"and one in five residents characterizing CBAs as a “bribe”. This approach is both didactic and exemplary for Black communities.

The Black masses must understand that Big Tech cartels are not in the business of improving material conditions for poor and working class Black communities - that’s what governments, pushed by the power of collective organizing and truly democratic processes and institutions, are supposed to do. This is yet another reason why the Black masses must direct the same antipathy to the Black Misleadership Class and the organizations and lawmakers within the Democratic Party like the CBC who uphold it, as those seeking to invade and pollute Black communities with hyper-scale data centers to power a technology that will only uphold and optimize oppressive systems like capitalism and white “supremacy.” The ra(AI)cial divide of AI and hyper-scale data center build-out does not have to be an inevitability. Instead, it can act as a clarion call that engenders a quintessential and generational organizing opportunity that can allow for a true confluence of poor and working class masses that confront and vanquish the Big Tech behemoth of late-stage capitalism while also fending off any possibility of a cross-class white “supremacy” consolidation. This, as veraciously suggested by writer Adryan Corcione, will require approaching the proliferation of AI and hyper-scale data centers with an anti-colonial perspective.

Anthony Karefa Rogers-Wright is a son of Sierra Leone, an international climate and environmental liberation advocate, a racial justice practitioner, a writer and policy expert residing in the United States His radio program, “Full Spectrum with Anthony Rogers-Wright,” airs on the Mighty WPFW network every Tuesday at 6:00 PM EST.

data centers

environmental racism

xAI

Artificial Intelligence

white supremacy

Congressional Black Caucus

Do you need and appreciate Black Agenda Report articles? Please click on the DONATE icon, and help us out, if you can.